492 ON IMPRESSIONS OF COLD 
pointed to the zenith, marked only five millesimal degrees ; 
but, on lowering it successively to the angle of 30 degrees 
above the horizon, it continued to indicate still the same effect. 
Water almost completely absorbs the pulsatory impressions of 
heat or cold ; and may not clouds, consisting of diffuse aque- 
ous particles, produce a similar effect ? But the feeble action 
of five degrees, amounting scarcely to the eight part of what is 
observed in clear weather, could not be any remnant of the 
pulses from the higher celestial regions, which had penetrated 
through the mass of vapours ; because, if the vertical transit, 
through the obstructing range, allowed only an eighth part to 
escape, the oblique passage of 30 degrees, redoubling the ex- 
tent of absorption, would have reduced the final discharge to 
five-eighths of a degree. The impression measured by the 
zthrioscope, in this case, must therefore have originated wholly 
in the strata of air between the under surface of the clouds 
and the ground. But in that narrow space, the extreme diffe- 
rence of temperature would be comparatively small. Hence 
the frigorific action is found always to diminish as the clouds 
descend. Nor does their variable denseness appear materially 
to affect the result, which is often the least, when a very thin, 
whitish, but low vapour, gathers in the atmosphere. Hence 
the ethrioscope might, with great facility, be employed in es- 
timating the altitude of clouds. 
As the higher strata of the atmosphere thus dart cold pulses 
downwards, so the lower strata must evidently project equal 
pulses of heat upwards. But to measure these, it would re- 
quire, as in fig. 7. the zthrioscope to be inverted and furnished 
with a pendant differential thermometer. The instrument, now 
carried to the top of a lofty mountain, and directed to the plain 
below, would indicate a considerable impression of heat, nearly 
proportional to the quantity of ascent ; and, therefore, amount- 
ing, for example, on the summit of Chimboraco, to perhaps 20 
millesimal 
